A real case: how to use micromobility data to decide where to install or relocate parking for bicycles and e-scooters.
Update: 🏆Turin is a winner of the SMAU Innovation Award thanks to this project.🏆 Read More
Executive Summary
-
Problem: Given the existing network of bike racks, there is a need to measure how efficiently racks are positioned and to identify areas with infrastructure shortages. Budgets and public space are limited, and the current network does not always capture demand.
-
Approach: Read the city through real micromobility usage data to see where vehicles accumulate and how often.
-
Method: City grid plus two simple indicators (Parking Index 1 = intensity of peaks, Parking Index 2 = frequency of accumulations) compared against the existing rack network.
-
Result: Four high priority areas emerged where the supply of parking is currently insufficient relative to real use.
-
Expected impact: Less disorderly parking, clearer sidewalks, better use of funds, transparent and defensible decisions.
Context: Why an Objective Method Was Needed
Like many European cities, Turin has invested in cycling and micromobility. Despite this, dedicated parking areas such as racks and marked bays are not always where they are needed. The outcome is:
-
vehicles left in sensitive points such as crosswalks, bus stops, and entrances;
-
conflict among users, residents, and shopkeepers;
-
requests to “install more” without a shared criterion.
Guiding Idea: Data Reveals Real Demand
Micromobility vehicles, both private and shared, move and stop in thousands of places every day. Where they stop more often is a proxy for parking demand. If we see repeated accumulations in an area, it is likely that orderly parking points are missing within that radius. Using these data, we can map real pressure and compare it with the current rack network to reveal gaps in coverage and redundancies in already well served zones or even oversupply.
Methodology, Explained Simply
Input Data
-
Positions and parking times of micromobility vehicles, aggregated and anonymized;
-
Map of existing bike racks, with location and capacity;
-
City grid, dividing the territory into regular cells for homogeneous comparisons.
The Tool: SWITCH Urbiverse
To avoid gut driven decisions, Urbiverse automatically analyzes parking patterns at the micro urban cell level and detects accumulations that signal unmet bicycle parking demand. The platform computes two comparable indices across the entire city (PI1 = intensity, PI2 = frequency), classifies critical cells parametrically, and produces maps, a gap analysis versus the existing network, and a prioritized list of interventions with suggested locations and recommended capacity. The workflow is repeatable, with configurable thresholds and layers, and is exportable via API or GIS to integrate with operations.
How Urbiverse Detects Accumulations
For each grid cell, Urbiverse processes parking and position data and computes two indicators normalized to the city baseline:
-
Parking Index 1, PI1 (intensity): how high the peak number of parked vehicles is at the same time compared with the rest of the city. Highlights areas with the worst crowding phenomena.
-
Parking Index 2, PI2 (frequency): how often the cell shows accumulations compared with the city. Highlights recurring hot spots.
Critical zone definition: by default, Urbiverse marks a cell as critical when PI1 > 0.5 and PI2 > 0.5, indicating strong and frequent accumulations. Thresholds are configurable according to the authority’s policy.
From Signal to Decision
-
Mapping critical cells
Urbiverse generates and updates a heat map of cells with high PI1 and PI2. -
Data fusion with what exists
The platform overlays the layer for existing bike racks from open data, municipal inventories, and surveys to provide a unified demand supply view. -
Index driven gap analysis
Urbiverse computes the mismatch between demand (PI1 and PI2) and supply (rack capacity and placement), producing a saturation index for each cell and an estimate of the expected benefit from new installations or upgrades. -
Operational prioritization
The platform returns an ordered list of micro areas of intervention with:-
suggested coordinates,
-
priority score,
-
estimated number of recommended parking spots,
-
reasons, that is, the features that influenced the score, for maximum transparency.
-
-
Ready to use outputs
Export of reports and maps, API and GIS integration, and interoperable formats for operations.
Technical note: to distinguish regular from episodic phenomena we use specific AI models, yet the output is a simple traffic light style map, with high, medium, and low classes, and a priority list to keep results interpretable and actionable.
What Emerged in Turin
-
Four areas show intense and frequent accumulations of micromobility vehicles.
-
In these areas rack density is low, so the current network does not capture demand.
-
The gap map provides a roadmap: start from these four zones, then extend to others.
Recurring Situations
-
Lunch break or office hours: many arrivals in a short time, disorderly parking on sidewalks.
-
School or university exit times: waves of arrivals with short but repeated peaks.
-
Interchange nodes for bus, tram, and metro: high turnover, constant demand for orderly bays.
Action Plan
Phase 1 – Data collection and alignment
-
Access provider data in basic formats, rack inventory, and policy criteria such as accessibility radius, aesthetics, safety, and constraints.
Phase 2 – Analysis and priority map
-
Compute PI1 and PI2, create the gap map, shortlist the first 10 to 15 candidate cells, and conduct targeted site visits.
Phase 3 – Pilot in 1 to 2 areas
-
Light installation using modular or temporary racks, or relocation of underused racks, plus neighborhood communication.
Phase 4 – Evaluation and scaling
-
Measure before and after, update the map, and extend to other priority areas.
Risks and Mitigations
-
Problem displacement: accumulations shift by 100 to 200 meters. Mitigation: continuous monitoring and micro adjustments.
-
Underuse of new racks: incorrect placement. Mitigation: start with modular installations and A/B test micro relocations.
-
Local opposition: concerns about space occupation. Mitigation: choose compact, orderly street furniture, share data and pilot duration, and present relocation alternatives.
Urbiverse All in One Intelligence
-
Hotspot maps for intensity and frequency;
-
Gap map overlaying demand and the current network;
-
Priority list with PI1 and PI2 justifications;
-
Guidelines for a 3 to 6 month pilot and scaling criteria.
Why Choose SWITCH
-
Integrated platform for advanced analytics, forecasting, and simulation in a single environment;
-
Operations in Italy and Europe;
-
Experience in mobility planning and operations projects.
Overall Results
With a few data types and a clear methodology, Turin obtained an objective basis for deciding where to install or relocate micromobility hubs to achieve a city that is tidier, more accessible, and transparent in its choices.
Want the same workflow for your project or company? Book a call or write to info@getswitch.io.
FAQ
How long for a first map?
Typically 2 to 4 weeks from data access.
Can we relocate existing racks instead of buying new ones?
Yes. The map also highlights redundancies. Where supply is excessive, relocation can be considered.
If I produce or install micromobility hubs such as smart racks, e-bike or e-scooter charging, or lockers, can I work with you?
Yes. With Urbiverse you can use the same data and signals that guided this case study to discover where to place your hubs. The platform automatically computes PI1 intensity and PI2 frequency of accumulations and returns maps and prioritized lists of the best micro areas based on your criteria.
Does the method also apply to e-scooters, not just bicycles?
Yes. It works with any mobility service that has trip data and designated parking areas.
Are the results usable in commercial offers and RFPs?
Yes. Outputs include exportable maps, tables, and dossiers that are useful for technical proposals, impact estimates, and installation planning.
How do you ensure privacy and compliance?
Data are processed in aggregated and anonymized form. No personal information is used.
Glossary
-
Sharing or micromobility: rental services for bicycles or e-scooters that anyone can pick up and leave in the city.
-
Proxy: a signal that represents a phenomenon. Here, vehicle parking stands represent parking demand.
-
Grid or cells: the territory divided into regular squares to analyze homogeneous areas.
-
Hotspot or accumulation: a point or zone where many vehicles concentrate.
-
Median: the value at the center of a set of numbers such that half of the numbers are above and half are below.
-
PI1, intensity: how large the maximum accumulation in an area was compared with the rest of the city.
-
PI2, frequency: how often accumulations repeat in that area compared with the city.
-
Gap analysis: comparison between where demand exists and where supply of racks is present.
-
Pilot: short period, small scale test to verify that the solution works before extending it.
